To use chat, please login.
Back to contacts

Strategist and EA Programmer

MartinC (MartinC)
Apr 04 2012 at 12:55
18 posts
Hi, I am new to this site so this is more of an introduction.

I have been programming EA's on demo accounts and live accounts for 6 months now.

For your information, those of you who program an EA and do a back test which reveals the mountains of pips your bot can collect... don't trust it. Back tests are different to live tests. And also, live demo tests are different again to live real money tests.

The environments are different and I've found this has different consequences.

At first, my EAs were running with huge profits on a backtest, even with over 95% accuracy on the test, but come to a live demo test the results were different. And then with a live money account even more different. You have to be careful when moving from demo to live. Don't be naive. Do your research and start with very low lots and increase as your confidence becomes more consolidated with the bots performance.

Different brokers can affect a bots performance also. I have experimented with many brokers and have settled on a good one which has very fast transactions and fast server ticks.

The first goal of an expert advisor is to reduce losses and make sure you will at least break even. With my EA, which has been tailored and refined over the last 6 months, I have managed to achieve that. I am now at the point of making 10% profit every day, with few days being in a minor loss. This is with demo accounts and live money accounts.

I was wondering if any one knew of any forex conferences which can be attended that will be good to meet people to share ideas and make connections. Or possibly, some established lectures or seminars which people attend to discuss EA programming.

My EA is based on an order flipping mechanism with a trailing stop loss and a time frame check. I won't reveal much more ;)

Good luck everyone with your EA programming!

Never leave till tomorrow what can be done today
raajfx (raajfx)
Apr 04 2012 at 19:35
4 posts
Hai, myself Raaj. Forex EA Creator from India.

Can contact me to Make New Forex Trading Robots(EAs) or Indicators as per your Ideas, to Modify Existing Robots as per your Needs and for other forex related software services.

raajfx@gmail.com
raajfx@ymail.com

mobile phone: (0091) 97 87 000 777

MartinC (MartinC)
Apr 05 2012 at 11:58
18 posts
Raaj, thanks for you advertisement but I am capable of EA programming myself.

Regards,
Martin

Never leave till tomorrow what can be done today
MetaCoder
Apr 07 2012 at 21:06
126 posts

   MartinC posted:
   Hi, I am new to this site so this is more of an introduction.

I have been programming EA's on demo accounts and live accounts for 6 months now.

For your information, those of you who program an EA and do a back test which reveals the mountains of pips your bot can collect... don't trust it. Back tests are different to live tests. And also, live demo tests are different again to live real money tests.


Your Wrong. Make a proper and extensive research before conclusion . 6 months is a very short time in the world of programming and Trading. FYI Backtest is very credible when done on Tick Data + variable spreads.

I suggest you do this. Run you EA for a period of say 2 weeks or more while logging tick data at the same time. After this period, run a backtest of your EA over the previous forward trading period - using the data you previously logged. You will obtain nearly an exact same backtest result.




My recommended Broker
PipGnostic
TheCyclist
Apr 11 2012 at 12:19
724 posts
Errrr MC,

Hate to break this to you, but run the same back-test on the same data on 2 different machines and you get a different result. Just the processors will change the result.

We tested back testing throughly and it's useless.

I've never matched a days live trading to a backtest on the same day. If MT is online or offline when you do a back test it will change the result. It makes up data if it can't find it.

I gave up on back tests years ago and just run live now days....

MetaCoder
Apr 11 2012 at 15:51
126 posts

   TheCyclist posted:
   Errrr MC,

Hate to break this to you, but run the same back-test on the same data on 2 different machines and you get a different result. Just the processors will change the result.

We tested back testing throughly and it's useless.

I've never matched a days live trading to a backtest on the same day. If MT is online or offline when you do a back test it will change the result. It makes up data if it can't find it.

I gave up on back tests years ago and just run live now days....


Are you aware that forward test on two different machines will yield different results as well ?

Secondly, Its your responsibility to provide tick by tick data for MT, else it'll make up data as your stated !

Also, indicators that make up an EA determines if the results must be considered or not. Williams FFM as an example amongst many others uses Fractal indicator which looks awesome on visual and MT backtest but fails miserable on forward test because fractals dissapears as soon as it has appeared on the chart and so on. My point is , indicators that make up an EA determines if its backtests must be considered or not.



My recommended Broker
MartinC (MartinC)
Apr 11 2012 at 20:46
18 posts
Yeah back tests are not reliable. So much can affect them.

If you want accurate results test your bots on a live demo account.

Never leave till tomorrow what can be done today
MartinC (MartinC)
Apr 11 2012 at 20:46
18 posts

   MetaCoder posted:
   

   MartinC posted:
   Hi, I am new to this site so this is more of an introduction.

I have been programming EA's on demo accounts and live accounts for 6 months now.

For your information, those of you who program an EA and do a back test which reveals the mountains of pips your bot can collect... don't trust it. Back tests are different to live tests. And also, live demo tests are different again to live real money tests.


Your Wrong. Make a proper and extensive research before conclusion . 6 months is a very short time in the world of programming and Trading. FYI Backtest is very credible when done on Tick Data + variable spreads.

I suggest you do this. Run you EA for a period of say 2 weeks or more while logging tick data at the same time. After this period, run a backtest of your EA over the previous forward trading period - using the data you previously logged. You will obtain nearly an exact same backtest result.






..and I've been programming for about ten years. Only six months with EA programming. EA programming is not difficult at all.

Never leave till tomorrow what can be done today
VinhLQ2
Apr 11 2012 at 20:48
3 posts
Hello,

I would like to build a simple EA as follows:


- EA to automatically enter both Sell and Buy orders at the same time, same lot size (entry price will be dynamically calculated based on spread by that time)
- I can change lot size at my wish
- I can preset SL and TP for both Sell and Buy order separately (set SL, TP in EA, and when orders are opened, SL and TP will be automatically added, SL, TP are in pips, working with both 4 and 5 demical system)
- I can set how many pips the next buy/sell orders will be triggered (for example, when market price moves up/down every 50 pips, a pair of buy/sell orders will be opened)
- I can set slippage
- Work on all currency pair, all brokers who allow EA on MT4
- I can set time, date to start and stop EA automatically

Thank you

Vincent

zero
sherryp
Apr 12 2012 at 06:25
28 posts

   MartinC posted:
   Hi, I am new to this site so this is more of an introduction.

I was wondering if any one knew of any forex conferences which can be attended that will be good to meet people to share ideas and make connections. Or possibly, some established lectures or seminars which people attend to discuss EA programming.

My EA is based on an order flipping mechanism with a trailing stop loss and a time frame check. I won't reveal much more ;)

Good luck everyone with your EA programming!


https://meetup.com

try searching for forex or trading.

good luck with your work.

sp.

Please login to comment .